from datetime import datetime
import pandas as pd
from pathlib import Path
import plotly
import plotly.express as px
import numpy as np
from statsmodels.tsa.api import VAR
import urllib.request
plotly.offline.init_notebook_mode()
NOW = datetime.now()
TODAY = NOW.date()
print('Aktualisiert:', NOW)
Aktualisiert: 2020-11-29 14:06:44.507268
STATE_NAMES = ['Burgenland', 'Kärnten', 'Niederösterreich',
'Oberösterreich', 'Salzburg', 'Steiermark',
'Tirol', 'Vorarlberg', 'Wien']
# TODO: Genauer recherchieren!
EVENTS = {'1. Lockdown': (np.datetime64('2020-03-20'), np.datetime64('2020-04-14'),
'red', 'inside top left'),
'1. Maskenpflicht': (np.datetime64('2020-03-30'), np.datetime64('2020-06-15'),
'yellow', 'inside bottom left'),
'2. Maskenpflicht': (np.datetime64('2020-07-24'), np.datetime64(TODAY),
'yellow', 'inside bottom left'),
'Soft Lockdown': (np.datetime64('2020-11-03'), np.datetime64('2020-11-17'),
'orange', 'inside top left'),
'2. Lockdown': (np.datetime64('2020-11-17'), np.datetime64(TODAY),
'red', 'inside top left')}
def load_data(URL, date_columns):
data_file = Path(URL).name
try:
# Only download the data if we don't have it, to avoid
# excessive server access during local development
with open(data_file):
print("Using local", data_file)
except FileNotFoundError:
print("Downloading", URL)
urllib.request.urlretrieve(URL, data_file)
return pd.read_csv(data_file, sep=';', parse_dates=date_columns, infer_datetime_format=True, dayfirst=True)
raw_data = load_data("https://covid19-dashboard.ages.at/data/CovidFaelle_Timeline.csv", [0])
additional_data = load_data("https://covid19-dashboard.ages.at/data/CovidFallzahlen.csv", [0, 2])
Downloading https://covid19-dashboard.ages.at/data/CovidFaelle_Timeline.csv Downloading https://covid19-dashboard.ages.at/data/CovidFallzahlen.csv
cases = raw_data.query("Bundesland == 'Österreich'")
cases.insert(0, 'AnzahlFaelle_avg7', cases.AnzahlFaelle7Tage / 7)
time = cases.Time
tests = additional_data.query("Bundesland == 'Alle'")
tests.insert(2, 'TagesTests', np.concatenate([[np.nan], np.diff(tests.TestGesamt)]))
tests.insert(3, 'TagesTests_avg7', np.concatenate([[np.nan] * 7, (tests.TestGesamt.values[7:] - tests.TestGesamt.values[:-7])/7]))
tests.insert(0, 'Time', tests.MeldeDatum)
fig = px.line(cases, x='Time', y=["AnzahlFaelle", "AnzahlFaelle_avg7"], log_y=True, title="Fallzahlen")
fig.add_scatter(x=tests.Time, y=tests.TagesTests, name='Tests')
for name, (begin, end, color, pos) in EVENTS.items():
fig.add_vrect(x0=begin, x1=end, name=name, fillcolor=color, opacity=0.2,
annotation={'text': name}, annotation_position=pos)
fig.show()
all_data = tests.merge(cases, on='Time', how='outer')
all_data.insert(1, 'PosRate', all_data.AnzahlFaelle / all_data.TagesTests)
all_data.insert(1, 'PosRate_avg7', all_data.AnzahlFaelle_avg7 / all_data.TagesTests_avg7)
fig = px.line(all_data, x='Time', y=['PosRate', 'PosRate_avg7'], log_y=False, title="Anteil Positiver Tests")
for name, (begin, end, color, pos) in EVENTS.items():
fig.add_vrect(x0=begin, x1=end, name=name, fillcolor=color, opacity=0.2,
annotation={'text': name}, annotation_position=pos)
fig.show()
states = []
rates = []
for state_name, state_data in raw_data.groupby('Bundesland'):
x = np.log2(state_data.AnzahlFaelle7Tage)
rate = 2**np.array(np.diff(x))
rates.append(rate)
states.append(state_name)
growth = pd.DataFrame({n: r for n, r in zip(states, rates)})
fig = px.line(growth, x=time[1:], y=STATE_NAMES, title='Wachstumsrate')
fig.update_layout(yaxis=dict(range=[0.25, 4]))
fig.show()
/usr/share/miniconda/lib/python3.8/site-packages/pandas/core/series.py:726: RuntimeWarning: divide by zero encountered in log2 /usr/share/miniconda/lib/python3.8/site-packages/numpy/lib/function_base.py:1280: RuntimeWarning: invalid value encountered in subtract
model = VAR(growth[150:][STATE_NAMES])
res = model.fit(1)
res.summary()
Summary of Regression Results
==================================
Model: VAR
Method: OLS
Date: Sun, 29, Nov, 2020
Time: 14:06:47
--------------------------------------------------------------------
No. of Equations: 9.00000 BIC: -42.9278
Nobs: 125.000 HQIC: -44.1369
Log likelihood: 1303.95 FPE: 2.97617e-20
AIC: -44.9641 Det(Omega_mle): 1.48883e-20
--------------------------------------------------------------------
Results for equation Burgenland
======================================================================================
coefficient std. error t-stat prob
--------------------------------------------------------------------------------------
const 0.600364 0.192487 3.119 0.002
L1.Burgenland 0.138473 0.089077 1.555 0.120
L1.Kärnten -0.310041 0.074616 -4.155 0.000
L1.Niederösterreich 0.037472 0.214508 0.175 0.861
L1.Oberösterreich 0.278432 0.176927 1.574 0.116
L1.Salzburg 0.139204 0.089675 1.552 0.121
L1.Steiermark 0.086828 0.126697 0.685 0.493
L1.Tirol 0.172546 0.083851 2.058 0.040
L1.Vorarlberg 0.021353 0.082085 0.260 0.795
L1.Wien -0.136031 0.168706 -0.806 0.420
======================================================================================
Results for equation Kärnten
======================================================================================
coefficient std. error t-stat prob
--------------------------------------------------------------------------------------
const 0.631203 0.248490 2.540 0.011
L1.Burgenland 0.001155 0.114995 0.010 0.992
L1.Kärnten 0.335018 0.096325 3.478 0.001
L1.Niederösterreich 0.104960 0.276919 0.379 0.705
L1.Oberösterreich -0.240394 0.228404 -1.052 0.293
L1.Salzburg 0.181130 0.115766 1.565 0.118
L1.Steiermark 0.234585 0.163560 1.434 0.152
L1.Tirol 0.135219 0.108248 1.249 0.212
L1.Vorarlberg 0.207788 0.105968 1.961 0.050
L1.Wien -0.564990 0.217791 -2.594 0.009
======================================================================================
Results for equation Niederösterreich
======================================================================================
coefficient std. error t-stat prob
--------------------------------------------------------------------------------------
const 0.323472 0.082540 3.919 0.000
L1.Burgenland 0.105465 0.038197 2.761 0.006
L1.Kärnten -0.026423 0.031996 -0.826 0.409
L1.Niederösterreich 0.135011 0.091983 1.468 0.142
L1.Oberösterreich 0.271155 0.075868 3.574 0.000
L1.Salzburg -0.008633 0.038453 -0.225 0.822
L1.Steiermark -0.059770 0.054329 -1.100 0.271
L1.Tirol 0.098163 0.035956 2.730 0.006
L1.Vorarlberg 0.150220 0.035199 4.268 0.000
L1.Wien 0.019539 0.072342 0.270 0.787
======================================================================================
Results for equation Oberösterreich
======================================================================================
coefficient std. error t-stat prob
--------------------------------------------------------------------------------------
const 0.199356 0.097886 2.037 0.042
L1.Burgenland 0.002782 0.045299 0.061 0.951
L1.Kärnten 0.031664 0.037945 0.834 0.404
L1.Niederösterreich 0.079117 0.109084 0.725 0.468
L1.Oberösterreich 0.355974 0.089973 3.956 0.000
L1.Salzburg 0.088209 0.045603 1.934 0.053
L1.Steiermark 0.195610 0.064430 3.036 0.002
L1.Tirol 0.029632 0.042641 0.695 0.487
L1.Vorarlberg 0.115224 0.041743 2.760 0.006
L1.Wien -0.098211 0.085793 -1.145 0.252
======================================================================================
Results for equation Salzburg
======================================================================================
coefficient std. error t-stat prob
--------------------------------------------------------------------------------------
const 0.743735 0.210492 3.533 0.000
L1.Burgenland 0.060438 0.097410 0.620 0.535
L1.Kärnten -0.014925 0.081595 -0.183 0.855
L1.Niederösterreich -0.075553 0.234573 -0.322 0.747
L1.Oberösterreich 0.065652 0.193477 0.339 0.734
L1.Salzburg 0.039291 0.098063 0.401 0.689
L1.Steiermark 0.115180 0.138549 0.831 0.406
L1.Tirol 0.219653 0.091695 2.395 0.017
L1.Vorarlberg 0.039401 0.089764 0.439 0.661
L1.Wien -0.166338 0.184487 -0.902 0.367
======================================================================================
Results for equation Steiermark
======================================================================================
coefficient std. error t-stat prob
--------------------------------------------------------------------------------------
const 0.219101 0.145386 1.507 0.132
L1.Burgenland -0.048455 0.067281 -0.720 0.471
L1.Kärnten -0.015349 0.056358 -0.272 0.785
L1.Niederösterreich 0.182978 0.162019 1.129 0.259
L1.Oberösterreich 0.394979 0.133634 2.956 0.003
L1.Salzburg -0.039056 0.067732 -0.577 0.564
L1.Steiermark -0.059387 0.095695 -0.621 0.535
L1.Tirol 0.202703 0.063333 3.201 0.001
L1.Vorarlberg 0.052665 0.061999 0.849 0.396
L1.Wien 0.122119 0.127424 0.958 0.338
======================================================================================
Results for equation Tirol
======================================================================================
coefficient std. error t-stat prob
--------------------------------------------------------------------------------------
const 0.282302 0.184433 1.531 0.126
L1.Burgenland 0.072712 0.085350 0.852 0.394
L1.Kärnten -0.083836 0.071494 -1.173 0.241
L1.Niederösterreich -0.126929 0.205533 -0.618 0.537
L1.Oberösterreich -0.106021 0.169524 -0.625 0.532
L1.Salzburg 0.001434 0.085923 0.017 0.987
L1.Steiermark 0.377904 0.121396 3.113 0.002
L1.Tirol 0.535469 0.080343 6.665 0.000
L1.Vorarlberg 0.233378 0.078651 2.967 0.003
L1.Wien -0.177743 0.161647 -1.100 0.272
======================================================================================
Results for equation Vorarlberg
======================================================================================
coefficient std. error t-stat prob
--------------------------------------------------------------------------------------
const 0.073216 0.212591 0.344 0.731
L1.Burgenland 0.029078 0.098381 0.296 0.768
L1.Kärnten -0.062858 0.082409 -0.763 0.446
L1.Niederösterreich 0.271004 0.236913 1.144 0.253
L1.Oberösterreich 0.022548 0.195407 0.115 0.908
L1.Salzburg 0.232646 0.099041 2.349 0.019
L1.Steiermark 0.160932 0.139931 1.150 0.250
L1.Tirol 0.048773 0.092610 0.527 0.598
L1.Vorarlberg 0.018641 0.090659 0.206 0.837
L1.Wien 0.214997 0.186327 1.154 0.249
======================================================================================
Results for equation Wien
======================================================================================
coefficient std. error t-stat prob
--------------------------------------------------------------------------------------
const 0.620460 0.117968 5.260 0.000
L1.Burgenland -0.007800 0.054592 -0.143 0.886
L1.Kärnten -0.003654 0.045729 -0.080 0.936
L1.Niederösterreich -0.058887 0.131464 -0.448 0.654
L1.Oberösterreich 0.274421 0.108432 2.531 0.011
L1.Salzburg 0.001704 0.054959 0.031 0.975
L1.Steiermark 0.010535 0.077648 0.136 0.892
L1.Tirol 0.076449 0.051389 1.488 0.137
L1.Vorarlberg 0.194653 0.050307 3.869 0.000
L1.Wien -0.098879 0.103394 -0.956 0.339
======================================================================================
Correlation matrix of residuals
Burgenland Kärnten Niederösterreich Oberösterreich Salzburg Steiermark Tirol Vorarlberg Wien
Burgenland 1.000000 0.095789 -0.045697 0.196653 0.240401 0.009924 0.065221 -0.124731 0.110787
Kärnten 0.095789 1.000000 -0.065035 0.175591 0.080663 -0.170699 0.190788 0.016714 0.267373
Niederösterreich -0.045697 -0.065035 1.000000 0.233831 0.067230 0.157163 0.065997 0.050050 0.356513
Oberösterreich 0.196653 0.175591 0.233831 1.000000 0.236702 0.266744 0.063388 0.055148 0.041895
Salzburg 0.240401 0.080663 0.067230 0.236702 1.000000 0.127093 0.043325 0.088466 -0.065963
Steiermark 0.009924 -0.170699 0.157163 0.266744 0.127093 1.000000 0.085693 0.083581 -0.196993
Tirol 0.065221 0.190788 0.065997 0.063388 0.043325 0.085693 1.000000 0.137493 0.089049
Vorarlberg -0.124731 0.016714 0.050050 0.055148 0.088466 0.083581 0.137493 1.000000 0.081017
Wien 0.110787 0.267373 0.356513 0.041895 -0.065963 -0.196993 0.089049 0.081017 1.000000